I am using cassandra 2.1.2 and datastax cassandra-driver-core-2.1.2. Here is a strange problem: When a keys (or table made, deleted) is created, then some customers received duplicated incidents, about 200+ times my clusters and my clients are at different places (one lane There is a lot of problems for this reason, once the customer receives such an event, then he should refresh the schema, and the system schema Infos has the system. Should be taken from the spaces and others. Finally, also refresh nodelist and tokenmap Due to all these actions some data may be transferred, and 200+ events are terrible in one second. So any body knows why & amp; How to stop?
Thank you for reading this.
When you mention "something that my clients receive duplicate incidents", I value I'm having that many cluster instances, is that right? If you have only one cluster object and you are getting many events, then I am thinking that this is a bug, the Java-driver has only 1 connection for schema, topology and node position changes according to the cluster frequency ('Control Connection' Nominated). This connection is initially established for one of your contact points (and if the connection is lost then it will select another node in the cluster).
Without understanding more about your configuration, I will consider the following:
- Follow, that is, only making 1 cluster frequency (JVM) per application .
- If you want to stop 1 node from being responsible for sending events to your clients, make your contact random, so the points are not always chosen primarily for the control connection. (Note: if there is a ticket for it then the Java-Driver can do this for you),
No comments:
Post a Comment